C A R B O N 8 6 ( 2 0 1 5 ) 3 1 8 – 3 2 4

.sc iencedi rect .com

Avai lab le at www

ScienceDirect

journal homepage: www.elsev ier .com/ locate /carbon

Role of residual polymer on chemical vapor growngraphene by Raman spectroscopy

http://dx.doi.org/10.1016/j.carbon.2015.01.0550008-6223/ 2015 Elsevier Ltd. All rights reserved.

* Corresponding author.E-mail address: syju@yonsei.ac.kr (S.-Y. Ju).

Eunhye Koo, Sang-Yong Ju *

Department of Chemistry, Yonsei University, Seoul 120-749, Republic of Korea

A R T I C L E I N F O

Article history:

Received 4 December 2014

Accepted 28 January 2015

Available online 7 February 2015

A B S T R A C T

Although various applications extensively utilize polymer-assisted graphene transfer step,

the role of residual polymer on graphene was not well-understood. Here, we report the

effect of poly (methyl methacrylate) (PMMA) on chemical vapor deposition-grown hexago-

nal graphene via Raman spectroscopy. Analysis of bare-, PMMA-covered supported, and

PMMA-covered suspended graphene exhibits that their G and 2D band positions are pro-

gressively downshifted in that order. Mapping of spatial G and 2D band shifts into doping

and strain contributions shows that PMMA residue exerts moderate 0.15% tensile strain on

graphene/substrate, as compared to that of bare graphene. During this tensile strain, resid-

ual PMMA-covered graphene maintains its doping level as much as bare graphene does.

2015 Elsevier Ltd. All rights reserved.

1. Introduction

Graphene, a one-atom-thick honeycomb lattice of sp2

hybridized carbon, has gathered immense attention due to

its exceptional electrical [1–3], mechanical [4], and optical

properties [5]. Its two-dimensional physico-chemical property

allows for versatile application in areas such as transparent

protective coatings [6,7], barrier films, and membranes [8,9].

Graphene has been prepared using various methods since it

was first isolated via mechanical exfoliation [10]: these

include chemical vapor deposition (CVD) using metal (nickel

[11], and copper [12]), SiC [13], and chemical exfoliation

[14–17]. Among those methods the CVD process provides

opportunity for large-area graphene as well as layer control

[8,12], which result in excellent electrical [18] and mechanical

[19] properties comparable to those of pristine exfoliated

graphene.

Graphene transfer on a desired silicon substrate utilizes

protective polymer layer such as poly (methyl methacrylate)

(PMMA) for various applications [8,11,20,21]. The PMMA

residue is known to affect electrical properties of graphene

[22] and the complete removal of PMMA has proven difficult

[23]. Therefore, the understanding of graphene in the pres-

ence of residual PMMA has a paramount importance. For this

matter, Raman spectroscopy is a powerful tool to decipher

this effect. Up to now, graphene containing residual PMMA

exhibits p-doping behavior upon electrical measurement

[24]: graphene with residual PMMA on substrate appears to

have upshifted G band and, upon thermal annealing,

PMMA-removed graphene has further upshifted G band from

that (1581 cm1) of charge-neutral monolayer graphene [25].

However, thermal annealing of graphene enhances its

p-doping and complicates to interpret the role of residual

PMMA. Recently, it has been shown that doping and strain

on graphene with underlying SiO2/Si substrate have different

trajectories of G and 2D band positions (xG and x2D,

respectively) from charge-neutral graphene, showing that

p- or n-doping has much slower Dx2D/Dx2D slope than strain

does [26]. This method allows us to probe accurate descrip-

tion of the residual PMMA role on graphene. In this study,

C A R B O N 8 6 ( 2 0 1 5 ) 3 1 8 – 3 2 4 319

we show that residual PMMA exerts mild tensile strain on

transferred CVD-grown graphene on substrate, probed by

Raman spectroscopy without any thermal treatment.

2. Experimental

2.1. Materials and instrumentations

All reagents are spectroscopic grade and used without further

purification. PMMA (molecular weight: 950 kDa, 2% dilution in

anisole (A2)) was purchased from MicroChem. Millipore qual-

ity deionized water with resistivity greater than 18 MX was

used for all experiments. Silicon substrates with 285-nm thick

SiO2 (Lot#:7400383-603-Z, degenerately p-doped, resistiv-

ity = 1–10X-cm, Shinetsu, Japan) were subject to Piranha

(H2SO4: H2O2 = 7:3) cleaning for 10 min and were washed with

copious amounts of water and dried with N2. Prior to use, the

substrates were stored in deionized water. CVD setup was

custom-made (see Fig. S1 in the Supplementary data (SD)

for the detailed setup). A pre-patterned marker and well were

obtained via conventional lithography techniques (see SD for

the detailed procedure). SEM images were acquired with

JEOL-6701F, with an acceleration voltage of 3 keV. FT-IR

spectra were collected using Vertex70 (Bruker), equipped with

a DLaTGS detector. The graphene-containing sample and

spin-coated PMMA on a Si substrate were placed into the

beam bath, and the FT-IR spectra were acquired by accumu-

lating 100 times. A control sample was also prepared by

spin-coating the PMMA solution on the Si substrate at

3000 rpm for 60 s to produce a 75-nm thick layer. Atomic force

microscopy (AFM) was measured using JPK nanowizard via

tapping mode. A silicon cantilever (force constant: 37 N/m,

ACTA-20, App Nano) was utilized with resonance frequency

at 300 kHz.

2.2. CVD growth of graphene

Graphene was grown using a custom-made low pressure

CVD (LP-CVD) setup. A copper strip (0.5 · 4 cm2, purity

>99.96%, Cu-113213, Nilaco Corp.) was initially dipped into

acetic acid for 10 min to remove oxide layer. The Cu foil

was washed with isopropanol and dried under an N2

stream (99.99%). The dried copper strip was immediately

placed in the hot zone of a fused silica tube 1 inch in

diameter, and 60 standard cubic centimeters (sccm) of H2

gas (99.9999%) was flowed onto the copper foil in the tube

up to 1000 C. Once the temperature reached 1000 C, a

mixture of 1 sccm CH4 (99.95%, Donga Gas) and 60 sccm

H2 was continuously flowed for 80 min at the same temper-

ature. Following the CVD growth of graphene, the stream

of 60 sccm of H2 was maintained while the sample was

slowly cooled to 230 C at a rate of 25 C/min. This

method corresponds to the samples in Figs. 1 and 3. Alter-

natively, when partial graphene overgrows, we reduced

flow speed by half for 60 min at 980 C, corresponding to

the samples in Fig. 2. In order to visualize the graphene

layer, the graphene-grown copper sample was optionally

oxidized at 165 C for 6 min on a hot plate in an ambient

atmosphere [27].

2.3. Graphene transfer onto a SiO2/Si substrate

The process of transferring graphene onto a desired substrate

follows that of previously-reported literature [8]. PMMA was

spin coated at 3000 rpm for 60 s onto a graphene-grown Cu

foil as a supporting layer. Following the optional oxidation,

the underlying copper foil was removed via submersion in a

0.1 M ammonium persulfate solution for 3 h, and the floating

graphene with the PMMA layer was scooped out with a clean

slide glass and subsequently washed with deionized water

twice. The copper-removed, PMMA-supported graphene was

gently transferred to a slightly-tilted Si substrate with a

285-nm thick SiO2 layer. After drying, the sample was further

dried in a vacuum oven under 30 mmHg at room temperature.

The protective PMMA layer was removed via dipping into an

acetone bath overnight. The substrate was gently treated with

sequential acetone and isopropanol washing, and was finally

dried with an N2 stream.

2.4. Optical/contrast measurement

An optical measurement was conducted using a backscatter-

ing geometry via either 20· (numerical aperture (NA): 0.40,

Olympus) or 100· (NA: 0.90, Olympus) objective lenses. Either

a CMOS camera (3.6 lm/pixel, 1280 · 1024, DCC1645C, Thorl-

abs) or a Si charge-coupled device (CCD) (6.45 lm/pixel,

1392 · 1040, CoolSNAP HQ2, Photometrics) with optional

550 nm Band pass filter (FWHM = 10 nm, FB550-10, Thorlabs)

to the emission path were used to measure relative and quan-

titative contrast, respectively.

2.5. Raman measurements

Micro-Raman spectroscopy was constructed using a home-

built upright microscope (BX-50, Olympus) as shown in

Fig. S2A, according to previous literature [28]. The slit of a

spectrometer (Triax320, 1800 gratings/mm, Horiba Jobin–

Yvon, see Fig. S2B spectrometer efficiency according to polar-

ization direction) with a Si array CCD (Symphony, 26 lm/

pixel, 1024 · 256, CCD-1024 · 256-BIDD-STE, Horiba Jobin–

Yvon) was placed at the focal length (180 mm) of a tube lens

in a trinocular piece (U-TR30, F.N.: 22, Olympus). A diode-

pumped solid state laser (532 nm, maximum power = 2 W)

was used as an excitation light source, which produces

vertically-polarized beam after passing consecutive a laser

cleaning filter (LL01-532-12.5, Semrock) and a Glan-Taylor

polarizer (GT5, Thorlabs). Objective lenses with 50· (Mplan,

NA = 0.75, Olympus), and 100· (Mplan, NA = 0.90, Olympus)

were used to focus and form a diffraction-limited spot. A

combination of two-stacked trinocular intermediates (U-

TRU, Olympus) installed with two desired filter sets further

controlled the switch between the Raman analysis and the

optical imaging: the bottom intermediate has two cubes con-

taining a dichroic long-band pass filter (LPD01-532RU-

25 · 36 · 1.1, Chroma) and a 50/50 beam splitter. The top inter-

mediate has a 50/50 beam splitter and long-band pass filter

(LP03-532RE-25, Chroma) with a side port connected to an

external collimated LED light source (white light, M625L3-L1,

Thorlabs). For wide-field Raman measurement, an optional

Fig. 2 – A bare supported HGG on a SiO2/Si substrate. (A) OM

image of a HGG with occasional PMMA residue shown in

black. Inset: the corresponding SEM image. The optical

contrast with scales (bottom) was obtained along the yellow

dashed line. (B) The corresponding line Raman spectra from

the D (left), G (middle), and 2D bands (right) using a 100·objective lens. The inset of (C) illustrates the Dirac cone with

p-doping. CB and VB stand for conduction and valence

bands, respectively. (A color version of this figure can be

viewed online.)

Fig. 1 – Existence of residual PMMA on graphene. OM images of HGG from (A) graphene on copper foil sample, and (B) the

corresponding transferred graphene on a SiO2/Si substrate after copper etching and PMMA removal with acetone. (A color

version of this figure can be viewed online.)

320 C A R B O N 8 6 ( 2 0 1 5 ) 3 1 8 – 3 2 4

convex lens (AC254-200-A-ML, f = 200 mm, doublet, Thorlabs)

and a beam expander (1–8·, Cat. #: 68479, Edmunds) were

introduced between the laser and a beam splitter to focus

the excitation light into the back focal plane of an objective

lens. Vertically-polarized beam parallel to edge of hexagonal

graphene was introduced to sample (Fig. S2A). Calibration of

Raman spectroscopy was conducted using multiple

Hg/Ar lamp peaks (365.015, 404.656, 435.833, 546.074, and

579.066 nm from an Hg lamp, and 800.616, 866.794, and

912.297 nm from Ar) using a light source (HG-1, Ocean Optics)

over the entire visible range. Third-order polynomials are

used to produce the wavelength linearity [26]. Si peak at

520.89 cm1 was used as an internal reference, as shown in

Fig. S3. Typical acquirement of micro Raman spectrum was

obtained using a laser beam power below 1.3 mW and an

accumulation time of 50 s to avoid laser-induced heating

[29]. As in a previously-reported study [28], WFR spectroscopy

can be switched from micro Raman by placing a convex lens

and beam expander in the excitation path. WFR spectroscopy

in the presence of the convex lens utilizes 39 mW laser power

using a 100· objective lens with a collection time of 500 and

300 s for imaging and line spectra, respectively. Each of the

WFR images was normalized to the background images of

the substrate used. As shown in Fig. S2, placing band-pass fil-

ters corresponding to 580 and 620 nm (FB580-10 and FB620-10,

respectively, FWHM = 10 nm, Thorlabs) in the emission path

enables imaging of the G and 2D bands, respectively. A CCD

array detector with a spectrometer can image the specific

Raman bands with grating angles at 0. By placing a kinematic

slit at the focal plane in front of the spectrometer and chang-

ing the grating angle, as shown in Fig. S4, the imaging mode is

converted to a line Raman spectrum mode, while the

employed band pass filter is removed from the emission

beam path. Typically, line Raman spectra whose center wave-

lengths are 547, 580, and 620 nm are collected for 300 s as Si

reference, G band, and 2D band, respectively. These spectra

were corrected against instrumental variations using a

correction factor. To induce defect, laser irradiation using

widefield illumination was applied with 50 mW for 10 h.

3. Results and discussion

Hexagonal graphene grain (HGG) was synthesized because it

possesses single-domain crystallinity with a zigzag-termi-

nated edge and minimizes the heterogeneous effect of poly-

crystalline graphene [23,30]. HGG growth was achieved with

a low carbon feedstock via the LP-CVD method using a copper

strip (0.5 · 4 cm2), according to prior method [28]. As shown in

Fig. 1A, optical microscope (OM) image clearly shows HGG

after oxidizing the underlying Cu strip at 160 C for 6 min

[27]. The average length of HGG was about 25 lm. The sample

was transferred on 285-nm thick SiO2/Si substrate using

75-nm thick PMMA layer. After overnight ‘‘removal’’ step of

Fig. 3 – A PMMA-covered, suspended HGG grain over a well. (A) An SEM image, and the corresponding WFR images of the (B) G

and (C) 2D bands. Scale bar: 3 lm. Representative Raman spectra of (D) supported, and (E) suspended graphene regions,

respectively. The numbers in-plane and in the bracket note the position and FWHM of the respective bands. (F) I2D/IG ratio,

and (G) FWHM (C) of the G and 2D bands according to the distance extracted from the yellow dashed line in (B). (H) Plot of xG

versus x2D according to the graphene regions. (A color version of this figure can be viewed online.)

C A R B O N 8 6 ( 2 0 1 5 ) 3 1 8 – 3 2 4 321

PMMA in acetone bath, the HGG sample (Fig. 1B) still exhibits

PMMA residue (indicated by arrow). Pale lines devoid of PMMA

are formed along Cu grain boundaries, originating from low

adhesion of spuncoated PMMA layer to the recessed Cu grain

boundary. The existence of PMMA was confirmed by

comparing Fourier transform (FT)-infrared (IR) spectrum of

the sample with that of bulk PMMA sample (Fig. S5A and B),

exhibiting identical PMMA fingerprint such as 1732 cm1 from

C@O stretching band (see SD for detailed explanation). Utiliz-

ing an exposed and PMMA-covered HGGs, WFR spectroscopy

[28] was conducted to obtain effect of PMMA on graphene

with spatial and spectral information.

Bare supported HGG was first examined. HGG sample with

10-lm wide sample is clearly visible in the OM image (Fig. 2A)

and scanning electron microscope (SEM) image (inset of

Fig. 2A). This sample was obtained from the PMMA-uncovered

region, indicated by red arrow in Fig. S5. Upon closer look,

although there are remaining PMMA residues as shown in

black speckle, most of HGG surface is exposed to air. In order

to confirm that this is monolayer graphene, we obtained aver-

age contrast value C, which is defined by C = (Ro R)/Ro,

where R and Ro are the reflectance of graphene/substrate

and substrate only, respectively (see Section 2). Average con-

trast value of PMMA-free graphene region (black trace of

Fig. 2A) exhibits 0.08, in good agreement with those (i.e.

0.09) obtained from mechanically-exfoliated single-layer

graphene (SLG) having the same SiO2 thickness (i.e., 285 nm)

[31]. It is noteworthy that PMMA residue also contributes to

contrast value on top of that of graphene. Spatial distribution

of Raman spectra was acquired using WFR along the yellow

dashed line (see Section 2 for the detailed line Raman mea-

surement). The synthesized monolayer HGG exhibits intense

2D band near 2679 cm1 compared to G band near 1591 cm1,

as shown in Fig. 2B. We observed spatial variation in G and 2D

band positions which did not correlate with D bands near

1345 cm1, suggesting the heterogeneity near local graphene

environment. The representative Raman spectrum (Fig. 2C)

exhibits G and 2D bands at 1591 and 2679 cm1, having inten-

sity ratio of 2D over G bands (I2D/IG) with 1.5. The reason for

upshifted G band position was previously ascribed to sub-

strate-induced O2 doping of the bare graphene [32]. As a

result, p-doping of Dirac cone at K point of Brillouin zone

occurs as shown in inset of Fig. 2C. Such p-doping changes

bond strength of C–C bonds and leads to simultaneous G

and 2D bands upshift. The optical transition in the doped

state is prohibited and results in G and 2D bands stiffening.

Since defect on graphene also affects G and 2D positions

[33], the degree of defect as a function of intensity of D over

G bands (ID/IG) was also probed. Fig. S6 exhibits line Raman

spectra of the same line after 10 h 532-nm laser irradiation

(see Section 2). Defect further upshifts xG and x2D by 5 and

3 cm1 when ID/IG is 50%, which will be considered in later

discussion.

Now, we probed graphene containing PMMA residue sus-

pended over a well because suspended, PMMA-covered

graphene displays phonon modes decoupled from a underly-

ing substrate. Fig. 3A shows a SEM image of PMMA-covered

HGG over a 6.7-lm wide well which was fabricated by conven-

tional photolithography step. The corresponding G and 2D

band images (Fig. 3B and C, respectively) using bandpass

322 C A R B O N 8 6 ( 2 0 1 5 ) 3 1 8 – 3 2 4

filters (10 nm bandwidth, see Section 2) display spatially-vary-

ing intensities (below 35% and 38% from the average values,

respectively) over entire region except in the well whose

intensity is much greater owing to the difference in reflectiv-

ity of the underlying substrate. The intensities of the 2D

bands (Fig. 3C) are nearly two times higher than those of

the G band (Fig. 3B) in the supporting region of the monolayer

HGG. Raman spectrum obtained from a supported region on

the substrate (Fig. 3D) exhibits G and 2D band positions

(FWHM) at 1588 (16) and 2671 (33) cm1, respectively, showing

an intensity of 2D over G band (I2D/IG) ratio of 1.9 with no

defect peak near 1345 cm1. The observed G (2D) band posi-

tion is 3 (8) cm1 downshifted from the bare monolayer

graphene in Fig. 2, showing much larger shift in x2D as com-

pared to xG. This suggests that PMMA residue makes graph-

ene approach to the charge neutral point. Furthermore, the

PMMA-covered, suspended graphene over well further down-

shift xG (x2D) by 0.5 (2.7) cm1 from the supported graphene,

suggesting that role of PMMA is significant on graphene.

Part F and G of Fig. 3 illustrate the spatial I2D/IG ratio and

the FWHM of the G and 2D band (CG and C2D, respectively)

trends. The average I2D/IG ratio increased from 2.0 to 2.6

from the supported to suspended regions, which can be

explained by the aforementioned phonon softening of G

band. In addition, while the average CG and C2D at the sup-

ported region are 16 and 33 cm1, respectively, those from

the suspended region exhibit increased widths by 1 and

3 cm1, respectively, supporting phonon-softening. The xG

and x2D (Fig. 3H) from the suspended region exhibits down-

shift from the PMMA-covered supported graphene whose

average xG and x2D are 1588 ± 1.1 and 2670 ± 1.6 cm1, respec-

tively. The average position change of the 2D band (Dx2D = 7

cm1) from supported to suspended region is more responsive

than that (DxG = 3 cm1) of the G band.

In the presence or absence of residual PMMA, changes in

position and width of G and 2D bands were significant. For

this, defect-containing Raman spectra (i.e., ID/IG > 20%) were

carefully discarded and the spatial Raman points from vari-

ous HGGs are presented in terms of respective band positions,

band width, and I2D/IG ratio (Fig. 4A–C, respectively. Mean val-

ues are tabulated in Table S1 to determine the physical nature

of Raman mode shifts [26]. Fig. 4A illustrates a plot of xG and

x2D extracted from various HGGs with different environments

(i.e., bare graphene/substrate, residual PMMA on graphene/

substrate, and residual PMMA on graphene suspended over

a well). xG and x2D values of charge-neutral suspended

graphene value (1581 and 2667.5 cm1, respectively) from

Ref. [34] are added after correction of energy dispersive x2D

by excitation shift factor (i.e., 88 cm1/eV) [35]. Using this as

a reference point, two Dx2D/DxG lines originating from either

large-sloped strain (eT, Dx2D/DxG = 2.2–2.8) or low-sloped

doping (en, Dx2D/DxG = 0.7) [26,36–38] along with scales are

drawn to distinguish main contribution of various HGG, as

shown in Fig. 4A. Mean value of bare graphene sample on

substrate positioned at 1st quadrant of two diagonal lines

with distribution of correlation points along strain line. This

suggests that bare CVD graphene undergoes concomitant p-

doping and compressive strain, contrasted with mechanically

exfoliated graphene having tensile-strain [39]. When graph-

ene is covered with residual PMMA, the average correlation

point of the graphene is positioned at 4th quadrant, having

p-doping along with tensile strain. Interestingly, doping level

(i.e., 5 · 1012 cm2) remains to that of bare graphene (Fig. 4B)

along strain line (i.e. 0.1%). In addition, the strain value

exerted by residual PMMA on graphene has approximately

+0.15% from the bare graphene sample, suggesting tensile

strain. The residual PMMA-containing graphene suspended

over well further downshifts the correlation mean position

along the strain line. Under this unrestricted condition,

PMMA exerted more tensile strain on graphene.

We illustrate the effect of residual PMMA shown in Fig. 4A

as schematics (Fig. 4B–D). Fig. 4B illustrates schematics of

effect of residual PMMA on graphene. Unlike mechanically-

exfoliated graphene [26], bare graphene conforms undulating

SiO2/Si substrate (top) and shows compressive strain. The

SiO2 substrate having height undulation with RMS roughness

of 0.1 nm (Fig. S7) seems to facilitate oxygen doping on graph-

ene, which exhibits pronounced p-doping concomitant with

strain. In the presence of residual PMMA (Fig. 4C), graphene

preferentially interacts with residual PMMA due to their sim-

ilar Hildebrand solubility parameters (i.e., 22.7 and 23 MPa1/2,

respectively) [41] as compared to that of SiO2. As a result,

graphene follows topography of residual PMMA, rather than

bare graphene does SiO2. In addition, because residual PMMA

on graphene has finite thickness over few nanometer [24],

graphene would less conform the substrate and is partially

suspended over the protrusions from the substrate (Fig. 4D).

The fact that doping value maintains in the presence/absence

of PMMA indicates that the contact between protruding SiO2

features and graphene is main contributor to doping and

PMMA has negligible effect. Similar trend has been observed

in single-walled carbon nanotube (SWNT) [42–44]: while

SWNT on quartz substrate does not have or weak photolumi-

nescence (PL) [43,45], SWNT embedded in PMMA [44] or on

sapphire [45] displays strong PL. This is attributed to signifi-

cant doping effect of SiO2 compared to neural PMMA or basic

sapphire. Suspended PMMA/graphene (Fig. 4D) over a well is

expected to sag due to the apparent increased weight, leading

to more tensile strain than that of supported case.

The FWHM of the G and 2D bands are plotted in Fig. 4E.

Increasing CG and C2D values were exhibited by bare-sup-

ported, defective, PMMA-covered supported, and PMMA-cov-

ered suspended HGGs in that order, with the average CG

(C2D) spanning from 17 (36) to 13 (31) cm1. The reason for line

broadening is that graphene under strain splits G bands into

low and high frequency modes, showing apparent mode

broadening. It is noted that broadening of C can stem from

two extreme cases: either phonon softening [46] due to C–C

bond order change from extremely clean graphene or broad-

ening due to defective graphene. In our observation, defective

graphene, whose ID/IG ratio is greater than 20%, shows lesser

broadening than that of PMMA-covered HGG does. Similar to

the strain slope of the correlation band position, the slope

DC2D/DCG (i.e., 0.59) between bare and PMMA-covered SLG

can be interpreted as broadening of the peak widths of the

G and 2D bands as a result of strain with maintained doping

level [46]. I2D/IG ratios from various HGG samples are also

abridged in Fig. 4F. PMMA-covered suspended HGG exhibits

a high average I2D/IG of 2.7, mainly due to the aforementioned

strain effect, followed by PMMA-covered supported HGG at 2,

Fig. 4 – Breaking strain and doping contributions from various HGGs. (A) Correlated plot of xG and x2D. The red and purple

dotted line indicates strain-induced (eT) and doping-induced (en) position changes, obtained from Refs. [40,26]. Diamond

symbol indicates charge-neutral graphene. (B–D) Schematic illustration of (B) bare HGG on SiO2/Si substrate, (C) residual

PMMA/HGG on the substrate, and (D) residual PMMA/HGG suspended over a well. (E) Plot of the G and 2D band FWHM, where

the red dotted line is drawn for guidance of strain line. (F) I2D/IG ratio according to various HGG samples. (A color version of

this figure can be viewed online.)

C A R B O N 8 6 ( 2 0 1 5 ) 3 1 8 – 3 2 4 323

and finally bare supported graphene with the lowest average

value of 1.5. The I2D/IG ratios are significantly affected by ten-

sile strain, and defects to a lesser extent. Those I2D/IG ratios

can be further calibrated with spectrometer efficiency accord-

ing to each wavelength (by factor of 1.06) and polarization

direction, as shown in Fig. S2A.

4. Conclusions

We have presented a systematic Raman study on synthesized

HGGs with diverse PMMA and substrate environments. Com-

plicate Raman G and 2D bands were analyzed without any

thermal treatment according to position, width, and I2D/IGratio. Mapping of the correlated G and 2D band positions

reveals strain and doping contributions from various HGGs.

Strong polymer-graphene interaction enables graphene

adhere to residual PMMA and exhibits moderate tensile

strain, as compared to compressive-strained bare graphene

on substrate. In addition, graphene covered by residual PMMA

maintains its doping level as similar as that of bare graphene,

suggesting protruding substrate feature as main doping con-

tributor. This study provides an explanation about role of

residual PMMA with underlying graphene/substrate in terms

of strain, doping, and defects.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

The authors appreciate J. Kim for help with the contrast

simulation. This research was supported by Basic Science

Research Program through the National Research Founda-

tion of Korea (NRF) funded by the Ministry of Education,

Science and Technology (2014R1A1A2055572), and in part

by the Yonsei University Future-Leading Research Initiative

of 2014.

Appendix A. Supplementary data

Supplementary data associated with this article can be found,

in the online version, at http://dx.doi.org/10.1016/j.carbon.

2015.01.055.

R E F E R E N C E S

[1] Novoselov KS, Geim AK, Morozov SV, Jiang D, Katsnelson MI,Grigorieva IV, et al. Two-dimensional gas of massless diracfermions in graphene. Nature 2005;438(7065):197–200.

[2] Zhang Y, Tan Y-W, Stormer HL, Kim P. Experimentalobservation of the quantum hall effect and berry’s phase ingraphene. Nature 2005;438(7065):201–4.

[3] Geim AK, Novoselov KS. The rise of graphene. Nat Mater2007;6(3):183–91.

[4] Lee C, Wei X, Kysar JW, Hone J. Measurement of the elasticproperties and intrinsic strength of monolayer graphene.Science 2008;321(5887):385–8.

[5] Nair RR, Blake P, Grigorenko AN, Novoselov KS, Booth TJ,Stauber T, et al. Fine structure constant defines visualtransparency of graphene. Science 2008;320(5881):1308.

[6] Chen S, Brown L, Levendorf M, Cai W, Ju S-Y, Edgeworth J,et al. Oxidation resistance of graphene-coated cu and cu/nialloy. ACS Nano 2011;5(2):1321–7.

[7] Bunch JS, Verbridge SS, Alden JS, van der Zande AM, ParpiaJM, Craighead HG, et al. Impermeable atomic membranesfrom graphene sheets. Nano Lett 2008;8(8):2458–62.

[8] Bae S, Kim H, Lee Y, Xu X, Park J-S, Zheng Y, et al. Roll-to-rollproduction of 30-inch graphene films for transparentelectrodes. Nat Nanotechnol 2010;5(8):574–8.

[9] Yuk JM, Park J, Ercius P, Kim K, Hellebusch DJ, Crommie MF,et al. High-resolution em of colloidal nanocrystal growthusing graphene liquid cells. Science 2012;336(6077):61–4.

324 C A R B O N 8 6 ( 2 0 1 5 ) 3 1 8 – 3 2 4

[10] Novoselov KS, Geim AK, Morozov SV, Jiang D, Zhang Y,Dubonos SV, et al. Electric field effect in atomically thincarbon films. Science 2004;306(5696):666–9.

[11] Kim KS, Zhao Y, Jang H, Lee SY, Kim JM, Kim KS, et al. Large-scale pattern growth of graphene films for stretchabletransparent electrodes. Nature 2009;457(7230):706–10.

[12] Li X, Cai W, An J, Kim S, Nah J, Yang D, et al. Large-areasynthesis of high-quality and uniform graphene films oncopper foils. Science 2009;324(5932):1312–4.

[13] Berger C, Song Z, Li T, Li X, Ogbazghi AY, Feng R, et al.Ultrathin epitaxial graphite: 2d electron gas properties and aroute toward graphene-based nanoelectronics. J Phys Chem B2004;108(52):19912–6.

[14] Hummers WS, Offeman RE. Preparation of graphitic oxide. JAm Chem Soc 1958;80(6):1339.

[15] Hernandez Y, Nicolosi V, Lotya M, Blighe FM, Sun ZY, De S,et al. High-yield production of graphene by liquid-phaseexfoliation of graphite. Nat Nanotechnol 2008;3(9):563–8.

[16] Lotya M, Hernandez Y, King PJ, Smith RJ, Nicolosi V, KarlssonLS, et al. Liquid phase production of graphene by exfoliationof graphite in surfactant/water solutions. J Am Chem Soc2009;131(10):3611–20.

[17] Green AA, Hersam MC. Solution phaseproduction ofgraphene with controlled thickness via densitydifferentiation. Nano Lett 2009;9(12):4031–6.

[18] Tsen AW, Brown L, Levendorf MP, Ghahari F, Huang PY,Havener RW, et al. Tailoring electrical transport across grainboundaries in polycrystalline graphene. Science2012;336(6085):1143–6.

[19] Ruiz-Vargas CS, Zhuang HL, Huang PY, van der Zande AM,Garg S, McEuen PL, et al. Softened elastic response andunzipping in chemical vapor deposition graphenemembranes. Nano Lett 2011;11(6):2259–63.

[20] Wang Y, Tong SW, Xu XF, Ozyilmaz B, Loh KP. Interfaceengineering of layer-by-layer stacked graphene anodes forhigh-performance organic solar cells. Adv Mater2011;23(13):1514–8.

[21] Lee WH, Park J, Sim SH, Jo SB, Kim KS, Hong BH, et al.Transparent flexible organic transistors based on monolayergraphene electrodes on plastic. Adv Mater 2011;23(15):1752–6.

[22] Chan J, Venugopal A, Pirkle A, McDonnell S, Hinojos D,Magnuson CW, et al. Reducing extrinsic performance-limiting factors in graphene grown by chemical vapordeposition. ACS Nano 2012;6(4):3224–9.

[23] Lin Y-C, Lu C-C, Yeh C-H, Jin C, Suenaga K, Chiu P-W.Graphene annealing: how clean can it be? Nano Lett2011;12(1):414–9.

[24] Suk JW, Lee WH, Lee J, Chou H, Piner RD, Hao Y, et al.Enhancement of the electrical properties of graphene grownby chemical vapor deposition via controlling the effects ofpolymer residue. Nano Lett 2013;13(4):1462–7.

[25] Pirkle A, Chan J, Venugopal A, Hinojos D, Magnuson CW,McDonnell S, et al. The effect of chemical residues on thephysical and electrical properties of chemical vapordeposited graphene transferred to sio2. Appl Phys Lett2011;99(12).

[26] Lee JE, Ahn G, Shim J, Lee YS, Ryu S. Optical separation ofmechanical strain from charge doping in graphene. NatCommun 2012;3:1024.

[27] Jia CC, Jiang JL, Gan L, Guo XF. Direct optical characterizationof graphene growth and domains on growth substrates. SciRep 2012;2.

[28] Havener RW, Ju S-Y, Brown L, Wang Z, Wojcik M, Ruiz-VargasCS, et al. High-throughput graphene imaging on arbitrarysubstrates with widefield Raman spectroscopy. ACS Nano2011;6(1):373–80.

[29] Cai W, Moore AL, Zhu Y, Li X, Chen S, Shi L, et al. Thermaltransport in suspended and supported monolayer graphenegrown by chemical vapor deposition. Nano Lett2010;10(5):1645–51.

[30] Vlassiouk I, Regmi M, Fulvio P, Dai S, Datskos P, Eres G, et al.Role of hydrogen in chemical vapor deposition growth oflarge single-crystal graphene. ACS Nano 2011;5(7):6069–76.

[31] Ni ZH, Wang HM, Kasim J, Fan HM, Yu T, Wu YH, et al.Graphene thickness determination using reflection andcontrast spectroscopy. Nano Lett 2007;7(9):2758–63.

[32] Ryu S, Liu L, Berciaud S, Yu Y-J, Liu H, Kim P, et al.Atmospheric oxygen binding and hole doping in deformedgraphene on a sio2 substrate. Nano Lett 2010;10(12):4944–51.

[33] Martins Ferreira EH, Moutinho MVO, Stavale F, Lucchese MM,Capaz RB, Achete CA, et al. Evolution of the Raman spectrafrom single-, few-, and many-layer graphene with increasingdisorder. Phys Rev B 2010;82(12):125429.

[34] Berciaud S, Ryu S, Brus LE, Heinz TF. Probing the intrinsicproperties of exfoliated graphene: Raman spectroscopy offree-standing monolayers. Nano Lett 2008;9(1):346–52.

[35] Mafra DL, Samsonidze G, Malard LM, Elias DC, Brant JC,Plentz F, et al. Determination of la and to phonon dispersionrelations of graphene near the dirac point by doubleresonance Raman scattering. Phys Rev B 2007;76(23):233407.

[36] Metzger C, Remi S, Liu M, Kusminskiy SV, Castro Neto AH,Swan AK, et al. Biaxial strain in graphene adhered to shallowdepressions. Nano Lett 2009;10(1):6–10.

[37] Ding F, Ji H, Chen Y, Herklotz A, Dorr K, Mei Y, et al.Stretchable graphene: a close look at fundamentalparameters through biaxial straining. Nano Lett2010;10(9):3453–8.

[38] Mohiuddin TMG, Lombardo A, Nair RR, Bonetti A, Savini G,Jalil R, et al. Uniaxial strain in graphene by Ramanspectroscopy: G peak splitting, gruneisen parameters, andsample orientation. Phys Rev B 2009;79(20):205433.

[39] Bao W, Miao F, Chen Z, Zhang H, Jang W, Dames C, et al.Controlled ripple texturing of suspended graphene andultrathin graphite membranes. Nat Nanotechnol2009;4(9):562–6.

[40] Das A, Chakraborty B, Piscanec S, Pisana S, Sood AK, FerrariAC. Phonon renormalization in doped bilayer graphene. PhysRev B 2009;79(15):155417.

[41] Hernandez Y, Lotya M, Rickard D, Bergin SD, Coleman JN.Measurement of multicomponent solubility parameters forgraphene facilitates solvent discovery. Langmuir2009;26:3208–13.

[42] Lee D, Ahn G, Ryu S. Two-dimensional water diffusion at agaphene–silica interface. J Am Chem Soc2014;136(18):6634–42.

[43] Xie L, Liu C, Zhang J, Zhang Y, Jiao L, Jiang L, et al.Photoluminescence recovery from single-walled carbonnanotubes on substrates. J Am Chem Soc2007;129(41):12382–3.

[44] Ai N, Walden-Newman W, Song Q, Kalliakos S, Strauf S.Suppression of blinking and enhanced exciton emission fromindividual carbon nanotubes. ACS Nano 2011;5(4):2664–70.

[45] Kiowski O, Lebedkin S, Hennrich F, Kappes MM. Single-walledcarbon nanotubes show stable emission and simplephotoluminescence spectra with weak excitation sidebandsat cryogenic temperatures. Phys Rev B 2007;76(7):075422.

[46] Huang M, Yan H, Chen C, Song D, Heinz TF, Hone J. Phononsoftening and crystallographic orientation of strainedgraphene studied by Raman spectroscopy. PNAS USA2009;106(18):7304–8.

- S1 -

Supplementary data for

Role of Residual Polymer on Chemical Vapor Grown Graphene by

Raman Spectroscopy

Eunhye Koo and Sang-Yong Ju*

Department of Chemistry, Yonsei University, Seoul 120-749, Republic of Korea

* Correspondence E-mail: syju@yonsei.ac.kr

Table of contents ......................................................................................................................................S1

Figure S1. Photograph of a custom CVD setup ........................................................................................S2

1. Photolithography of Si substrate ........................................................................................................S2

Figure S2. (A) Schematic optical setup of micro/wide-field Raman spectroscope and (B) Polarization

dependency of the grating ..........................................................................................................................S3

Figure S3. Slit size and laser power dependence of 520.89 cm-1 Si peak .................................................S4

Figure S4. Schematic procedure for wide-field Raman images and line Raman spectra .........................S4

2. IR comparison ......................................................................................................................................S5

Figure S5. (A) OM image of PMMA-free region from PGGs on a substrate and (B) Comparison of

fourier transform (FT)-IR spectra depending on the existence of PMMA ................................................S5

Figure S6. Line Raman spectra and plot of G (ωG) vs. 2D band positions (ω2D) from PGG in Fig. 2 after

laser irradiation ..........................................................................................................................................S6

Table S1. Average positions, FWHMs, and I2D/IG ratios obtained from various graphene grains ..........S6

Figure S7. (A) AFM height topography of a bare 285-nm thick SiO2/Si substrate. (B) The corresponding

height variations according to positions. ...................................................................................................S7

- S2 -

Figure S1. Photograph of custom-made CVD setup.

1. Photolithography of Si substrate.

4-inch diameter, 285 nm-thick SiO2/Si substrate (Lot#7400383-603-Z, degenerately p–doped,

resistivity = 1-10 Ω-cm, Shinetsu, Japan) was utilized throughout this study unless otherwise noted. The

photoresist (PR, Cat. #: AZ GXR601, AZ Electronic Materials) was spun-coated at 1500 rpm for 85 sec

on the Si substrate to produce 1.2 µm-thick photoresist layer. The photoresist-covered Si substrate was

post-baked for 90 sec at 110 °C and subsequently patterned by a contact mask aligner (MA6, Suss

MacroTec, Germany). The patterned substrate was further developed with a diluted

tetramethylammonium hydroxide (TMAH) solution (MIF300, AZ Electronic Materials). Reactive ion

etcher (RB4510, LAM Research) using fluorinated gas was utilized to etch away SiO2 on Si substrate

with desired hole depth.

- S3 -

Figure S2. (A) Schematic optical setup of micro/wide-field Raman spectroscope. The inset shows a photograph of the installed Raman setup. The green and red lines represent excitation and emission paths, respectively. The green arrow indicates vertical polarization direction. Notation for abbreviation: P: polarizer, CV: convex lens, DBS: dichroic beam splitter, BFP: back focal plane, TL: tube lens, and BPF: band pass filter. Optical components with a dashed box are removed for micro Raman measurement. (B) Polarization dependency of the grating used in this study.

- S4 -

Figure S3. (A) Kinematic slit size dependence with a laser power of 1.7 mW, and (B) laser power dependence of the 520.89 cm-1 Si peak.

Figure S4. Schematic procedure for obtaining (A) wide-field Raman images, (B) slit narrowing operated by 0° grating angle, and (C) the corresponding line Raman spectra with appropriate grating angle along the slit. The green arrow in (A) indicates polarization direction.

- S5 -

2. IR comparison

IR spectra were acquired from three samples: PMMA spin-coated on a SiO2/Si layer 285-nm in

thickness, residual PMMA on graphene on the substrate, and the bare substrate. IR spectra from spin-

coated PMMA on the substrate displays vibrational bands at 2928, 2861, 1732, 1448, and 1375 cm-1.

Bands at 2928 and 2861 cm-1 originate from the symmetric and asymmetric C−H stretching mode of the

−CH2− group. The 1732 cm-1 band is due to the stretching vibration of the C=O ester group. The 1448

and 1375 cm-1 bands are C−H bending modes of the –CH3 group. As a whole, the aforementioned

assignment is attributed to PMMA. These PMMA signals are obtained from the residual

PMMA/graphene/substrate sample, while the control substrate does not exhibit any IR features

originating from PMMA.

Figure S5. (A) OM image of the PMMA-free region from PGGs on a substrate. The red arrow indicates the graphene hexagon from Fig. 2. (B) Comparison of fourier transform (FT)-IR spectra of spin-coated PMMA on a SiO2/Si substrate (red trace), residual PMMA/graphene on a substrate after acetone washing with copious amounts (black trace), and bare silicon substrate (green trace). Each spectrum is offset from top spectra for clarity.

- S6 -

Figure S6. PMMA-free defective hexagonal graphene grain on a substrate obtained via prolonged laser

irradiation. (A) The corresponding line Raman spectrum from the D (left), G (middle) and 2D bands (right). Note that the left side has color-coded arrows for the following Raman spectra. (B) The respective defect-derived Raman spectra. (C) Plot of ωG vs. ω2D in terms of increasing ID/IG values with 0.1 intervals. Standard deviation is presented with the exception of 0.4−0.6 ID/IG ratios, whose data points are less than 3.

Table S1. Average positions, FWHMs, and I2D/IG ratios obtained from various graphene grains.

Types of graphene grain ωG (cm-1) ω2D (cm-1) ГG (cm-1) Г2D (cm-1) I2D/IG

PMMA-covered suspended 1587.2 ±0.7 2667.7 ±0.9 17.4 ±1.9 35.8 ±2.1 2.6 ±0.5

PMMA-covered supported 1587.7 ±1.1 2670.4 ±1.6 16.4 ±1.2 33.0 ±1.8 2.0 ±0.1

Bare supported 1591.2 ±1.1 2678.1 ±3.2 12.5 ±2.7 30.7 ±3.4 1.5 ±0.2

Defective (ID/IG > 20 %) 1593.3 ±1.7 2681.4 ±2.0 14.9 ±2.2 31.2 ±2.3 1.5 ±0.2

- S7 -

Figure S7. (A) AFM height topography of a bare 285-nm thick SiO2/Si substrate. (B) The corresponding height variations according to positions.